Format and availability of data
For a data-model comparison, the proxyreconstruction data needs first to be combined into one large data set with wide
temporal and spatial coverage. This requires detailed information on sampling
methods, age models and representativeness of proxy data. Fundamental to the
continuation of the model-data comparison effort is the availability of and accessibility to data sets of proxy reconstructions
and model output (Dittert et al., p. 30).
Many recent data-modeling efforts have
utilized proxy data sets made available in
data archives.
Scientific education—bridging
the gaps between disciplines
Earth System science is traditionally split
into various disciplines and sub-disciplines. Overall, the diversity of expertise
provides a solid base for interdisciplinary
research. However, to gain holistic insights into the Earth System requires the
integration of observations, paleoclimate
data and climate modeling. These different approaches of Earth System science
are rooted in the different disciplines (geology, physics, meteorology, oceanography, etc.), which cut across a broad range
of timescales. It is therefore necessary to
link these disciplines at a relatively early
stage in MSc/PhD programs. The linking
of data and modeling would enable graduate students from a variety of disciplines
to cooperate and exchange views on the
common theme of Earth System science,
and lead to a better understanding of local processes within a global context.
Computational and conceptual models of the Earth System provide the ability
to investigate different scenarios in biogeochemistry, such as the carbon cycle,
the structure of marine sediments, and
isotope distribution in climate components. Statistical analysis further provides
a synthesis, comparison and interpretation of paleoclimate and simulated data.
Training and education, particularly in
time-series analysis, data exploration,
process understanding and model interpretation, should all be key components
of future education.
Special Section: Data-Model Comparison
et al., p. 24). Models can also be tested
against past and present scenarios (van
Oldenborgh, p. 26; Mudelsee and Girardin, p. 28), which again helps to improve
climate predictability.
Northern hemisphere atmospheric blocking in ice core
accumulation records from northern Greenland
Norel Rimbu, G. Lohmann and K. Grosfeld
Alfred Wegener Institute for Polar and Marine Research, Bremerhaven, Germany; [email protected]
Atmospheric blocking is a large-scale, midlatitude atmospheric phenomenon often
associated with persistent quasi-stationary, synoptic-scale, high-pressure systems.
The formation, maintenance and collapse
of atmospheric blocking cause large-scale
circulation anomalies and strongly impact
weather patterns. Therefore, blocking regimes constitute a significant climatological feature.
Northern hemisphere blocking shows
important variability at different timescales. Blocking frequencies have shown
a downward (upward) trend over Atlantic
and European (west Pacific) sectors. Su-
perimposed on these linear trends, blocking frequencies show significant interannual and decadal variation (Barriopedro
et al., 2006). However, in these studies,
the blocking variability was derived over
the relatively short time period covered
by observational data with daily resolution. Here, we present the first attempt
to directly relate interannual and decadal variability of several high-resolution
snow accumulation records from northern
Greenland with northern hemisphere atmospheric blocking. First, we investigate
the relationship between atmospheric
blocking and accumulation variability dur-
ing the period covered by both accumulation and high-resolution (daily) observational data. Based on this relationship, we
then discuss the blocking variability over
the last 400 yr based only on the accumulation variability.
As an example of blocking circulation,
we present in Figure 1a the 500-hPa geopotential height (shaded) and horizontal
wind (vectors) during the mature phase
(3 February 1975) of the blocking event
that occurred in the North Atlantic sector
from 28 January to 9 February 1975 (Diao
et al., 2006). It can be seen that large-scale
westerly flow, which is strongly blocked
Figure 1: a) The 500-hPa geopotential height (shaded; m) and horizontal wind (vector; m/s) during the mature stage (3 February 1975) of the 28 January to 9 February 1975
blocking event that occurred in the North Atlantic region. Location of core sites (filled white circles) and EOF1 values of snow accumulation records from Greenland are also
indicated. b) The corresponding time coefficients (PC1) and the mean snow accumulation of the five records represented in a). The PC1 was scaled to have the same mean
and variance as the mean accumulation time series (mm w.e./yr = mm water equivalent per year).
5 PAGES News • Vol.16 • No 2 • April 2008
Special Section: Data-Model Comparison
over the northeast Atlantic, influences the
Greenland region.
The Alfred Wegener Institute for Polar and Marine Research (AWI) in Bremerhaven, Germany performed a North
Greenland Traverse (NGT) between 1993
and 1995 (Schwager, 2000). As a proxy
for the blocking variability, we analyzed
the yearly snow accumulation time series from five ice cores drilled during the
NGT (Fig. 1a): B16 (73.9°N, 37.6°W; Miller
and Schwager, 2004), B18 (76.6°N, 36.4°W;
Miller and Schwager, 2000a), B21 (80.4°N,
41.1°W; Miller and Schwager, 2000b), B26
(77.3°N, 49.2°W; Miller and Schwager,
2000c) and B29 (76.0°N, 43.5°W; Miller and
Schwager, 2000d). These shallow ice cores
reach depths up to 150 m. Mean accumulation rates vary between 104 and 179
mm water equivalent/yr. The period covered by all five annual resolution records
is 1502 to 1992 (or 491 years). These accumulation records are available through the
PANGAEA online environmental database
(www.pangaea.de).
Accumulation records from Greenland
are affected by both meteorological and
glaciological noise (Crüger et al., 2004).
To filter out this noise and to obtain the
dominant patterns of accumulation variability, an Empirical Orthogonal Function
(EOF) is applied to normalized accumulation anomalies over the entire 1502 to
1992 period. The first EOF (Fig. 1a), which
describes about 24% of variance, captures
in-phase variability of all accumulation records. Its associated time coefficients (PC1)
show variations very similar to the mean
accumulation of the five records (Fig. 1b).
Since blocking heights affect the
middle and upper troposphere, the 500
hPa geopotential anomaly field is widely
used to identify blocking events. A popular algorithm to identify blocking events
in this field, also used in our study, was
developed by Tibaldi and Molteni (1990).
For each longitude, the northern (southern) meridional 500 hPa geopotential
height gradients are calculated between
three base latitudes, 55°N, 60°N and 65°N
and the corresponding 20°N (S) latitudes.
A given longitude is defined as blocked
at a specific instant in time if at least one
gradient between the base and southern
latitudes is greater than zero, and at least
one gradient between northern latitudes
and the corresponding basic latitudes is
smaller than -10 m/degree latitude (Tibaldi and Molteni, 1990). The frequency of
blocking (with neither time persistence
or spatial criteria yet imposed) at a certain
longitude is defined as the ratio between
the number of days when that longitude
Figure 2: a) Blocking-like frequency as a function of longitude for years of high and low snow accumulation on
Greenland during the 1948-1992 period. Shaded area outlines the 5°W-5°E area discussed in the text. b) Frequency
of daily atmospheric patterns during high and low accumulation years during the period 1850-1992 grouped in
9 clusters according to Philipp et al. (2007). The arrow indicates the difference between frequency of cluster 3 for
high and low accumulation years.
was blocked to the total number of days of
the analyzed period.
The frequency of blocking events during high accumulation years (PC1 higher
than 0.75 standard deviation) is significantly higher in the Atlantic-European sector (20°W-100°E) than the corresponding
frequency during low accumulation years
(PC1 lower than -0.75 standard deviation).
A reverse situation is found for the sector
80°W-20°W (Fig. 2a). No significant differences between the frequencies of blocking
events during high and low accumulation
years were detected outside the Atlantic
and European sectors (Rimbu et al., 2007).
A cluster analysis of a daily-reconstructed
atmospheric circulation since 1850 revealed that winter atmospheric circulation
in the Atlantic-European region is optimally represented by 9 clusters (Philipp et
6 PAGES News • Vol.16 • No 2 • April 2008
al., 2007). To better assess the relationship
between the snow accumulation variability on Greenland and Atlantic-European
atmospheric circulation, we compared the
frequency of these clusters during high
and low accumulation years for the period
1850-1992 (Fig. 2b). Significantly higher
frequency of cluster 3 is recorded during
high relative to low accumulation years.
Cluster 3 represents a blocking high centered near the UK (Philipp et al., 2007).
A blocking-like circulation is associated with a blocking event if it is persistent in time and extends over several longitudes. The typical duration of blocking
episodes varies between 5 and 30 days
(Tibaldi and Molteni, 1990). Based on the
frequency distribution of blocking-like
circulation states, as represented in Figure
2a, we define a blocking index as the num-
and Oort, 1992) in the Atlantic European
region during years characterized by high
and low values of the above-defined blocking index. The axis of maximum moisture
transport, defined as vertically integrated
water vapor transport (vectors), shifts to
a more northeast-directed orientation
across the Atlantic and extends northward
to Greenland during years of high blocking index values (Fig. 3a) relative to the
times of low blocking index (Fig. 3b). A
significant reduction in the magnitude of
atmospheric moisture transport (contour
lines) over Greenland during periods of
less frequent blocking relative to periods
of frequent blocking is also evident.
We have shown that over periods covered by observed and reconstructed daily
atmospheric circulation data, high accumulation in our Greenland ice cores is associated with high blocking frequency in
the Atlantic-European region. Assuming
that this relationship is stable in time, the
periods of relatively high (low) accumulation during the last 400 years are characterized by high (low) blocking activity in
the Euro-Atlantic region. We detected significant peaks at bi-decadal (~20 yr) and
multi-decadal (~70 yr) timescales in the
PC1 of accumulation records, which may
also characterize the blocking activity in
the Atlantic region. Our analysis should be
extended to other proxy data, as well as to
climate model experiments, for a better
understanding of blocking variability during past periods.
Special Section: Data-Model Comparison
ber of days in a winter when in sector 5°W5°E (Fig. 2a, shaded region) at least three
consecutive longitudes are blocked for at
least four consecutive days. The blocking
index is significantly correlated with accumulation PC1 over the period 1948-1992.
This result is confirmed by a composite
analysis. It reveals that this sector was
blocked on average 13.6 days per winter
during high accumulation years, which is
significantly more frequent than the 8.6
days per winter during low accumulation
years.
To identify the physical mechanism
responsible for the relationship between
blocking activity and the dominant pattern of accumulation variability, we investigated the moisture transport (Peixoto
References
Diao, Y., Li, J. and Luo, D., 2006: A new blocking index and its application:
blocking action in the Northern Hemisphere, Journal of Climatology, 19: 4819-4839.
Philipp, A., Della-Marta, P.M., Wanner, H., Fereday, D.R., Jones, P.D. and
Moberg, A., 2007: Long-term variability of daily North AtlanticEuropean pressure patterns since 1850 classified by simulated
annealing clustering, Journal of Climate, 20: 4065-4095.
Rimbu, N., Lohmann, G. and Grosfeld, K., 2007: Northern Hemisphere
atmospheric blocking in ice core accumulation records from
northern Greenland, Geophysical Research Letters, 34: L09704,
doi:10.1029/2006GL029175.
Schwager, M., 2000: Ice core analysis on the spatial and temporal variability of temperature and precipitation during the late Holocene
in North Greenland, Reports on Polar Research, 362: Alfred
Wegener Institute for Polar and Marine Research, Bremerhaven,
Germany.
Tibaldi, S. and Molteni, F., 1990: On the operational predictability of
blocking, Tellus series A, 42: 343-365.
For full references please consult:
www.pages-igbp.org/products/newsletter/ref2008_2.html
Figure 3: Water vapor transport (vector) during a) high and b) low values of blocking index (see text for definition).
The magnitude of the water vapor transport (kg/ms) is indicated by both contour lines and length of vectors.
7 PAGES News • Vol.16 • No 2 • April 2008